Vector trellis coded modulation using vector convolutional codes for reliable data transmission
A vector trellis coded modulation scheme (VTCM) is accomplished by blocking input samples into a sequence of input vectors of length L, vector convolutional coding the input vectors to map K input vectors into N output vectors and modulating the output vectors into symbols from an expanded alphabet. In a first case, each output vector is modulated to a different symbol thereby improving coding gain while maintain bandwidth efficiency. In a second case, blocks of N output vectors are modulated to one symbol thereby improving bandwidth efficiency by a factor of N while maintain coding gain. A polyphase/multirate representation of the vector convolutional codes is preferably used to generated the vector convolutional codes. In general, a computer search can be employed to find the polyphase coefficients that define the set of vector convolutional codes. When the input and output vectors have the same length, known scalar-valued convolutional codes are blocked to generate the vector convolutional codes. This provides good modulation codes without having to perform a computer search.
Latest Hughes Electronics Corporation Patents:
- Method and system for providing connection handling
- Proxy architecture for providing quality of service(QoS) reservations
- Method and system for providing differentiated quality of service diversity ALOHA
- Apparatus, method, and system for removing ethernet header and adding second IP header
- METHOD AND APPARATUS FOR ADAPTIVE CHANNEL SELECTION
Claims
3. The method of claim 2, wherein the input and output vectors have the same length M, said N M.times.M matrix-valued filter transfer functions R.sub.0 (z), R.sub.1 (z)... R.sub.N-1 (z) being generated by:
- providing scalar-valued filter transfer functions H.sub.0 (z), H.sub.1 (z),... H.sub.N-1 (z) for a K/N scalar convolutional code;
- computing impulse responses h.sub.0 (z), h.sub.1 (z),... h.sub.N-1 (z) for the scalar-valued filter transfer functions;
- computing polyphase components Q.sub.ij for 0.ltoreq.i.ltoreq.N-1, 0.ltoreq.j.ltoreq.M-1 from the impulse responses; and
- arranging the M polyphase components Q.sub.ij for each matrix-valued filter transfer function into a pseudo-circulant matrix.
10. The method of claim 9, wherein the scalar-valued filter transfer functions are blocked into the pseudo-circulant matrices by:
- computing impulse responses h.sub.0 (z), h.sub.1 (z),... h.sub.N-1 (z) for the scalar-valued filter transfer functions;
- computing polyphase components Q.sub.ij for 0.ltoreq.i.ltoreq.N-1, 0.ltoreq.j.ltoreq.M-1 from the impulse responses; and
- arranging the M polyphase components Q.sub.ij for each matrix-valued filter transfer function into the pseudo-circulant matrix.
4922507 | May 1, 1990 | Simon et al. |
5023889 | June 11, 1991 | Divsalar et al. |
5321725 | June 14, 1994 | Paik et al. |
5395518 | March 7, 1995 | How |
5717706 | February 10, 1998 | Ikeda |
- P.P. Vaidyanathan, Multirate Digital Filters, Filter Banks, Polyphase Networks, and Applications: A Tutorial, Proceedings of the IEEE, vol. 78, No. 1, Jan. 1990, pp. 56-93. R. Johannesson et al., "Further Results on Binary Convolutional Codes with an Optimum Distance Profile", IEEE Transactions on Information Theory, IT-24, Mar. 1978, pp. 264-268. Shu Lin et al., "Convolutional Codes", Error Control Coding: Fundamentals and Applications, Chapter 10, Prentice-Hall, New Jersey, 1983, pp. 287-314. E. Biglien et al., TCM: Combined Modulation and Coding, Introduction to Trellis-Coded Modulation with Applications, Chapter M, Macmillan Publishing Company, New York, 1991, pp. 67-98.
Type: Grant
Filed: Jun 21, 1996
Date of Patent: Sep 15, 1998
Assignee: Hughes Electronics Corporation (El Segundo, CA)
Inventor: Xiang-Gen Xia (Westlake Village, CA)
Primary Examiner: Chi H. Pham
Assistant Examiner: Bayard Emmanuel
Attorneys: V. D. Duraiswamy, W. K. Denson-Low
Application Number: 8/673,715
International Classification: H04L 2302; H04L 512;